TASA HUB - AI SAAS PRODUCT

Impact: Established and retained customer engagement as a start up. Beat standard speed of go to market compared to traditional SaaS 0-1 without the use of AI models.

Outcome: Launched as a fully functioning SaaS platform. From “We have an idea” → “We have customers using this”. 0-1 build using AI tools: Cursor AI, Lovable, Gemini and ChatGPT. As well as full API integrations.

I AM THE FOUNDING UX and UI PRODUCT DESIGNER FOR BOTH MOBILE AND WEB EXPERIENCES.

As an individual contributor, I lead by example with my partners and design peers.

Designing an AI-Powered SaaS Platform from 0 → 1

Role: Founding Product Designer
Team: 2 Founders, 1 Lead Engineer, Me
Timeline: 2024–2025
Tools: Figma, Cursor AI, ChatGPT, Gemini, Lovable, Supabase, Stripe, PostHog

The Challenge

TasaHub started as an idea. There was no product, no users, no workflows, and no validated problem — just founders with a hypothesis. My job was to turn ambiguity into a real product.

“The $ 72B Coffee and Boba industry for SMB lack the Enterprise AI layer that current POS Softwares don’t invest in. They are not giving the Coffee and Boba industry justice — even with SMB’s small margin.”
— Founders

What I Owned

As the Founding Designer, I led the product end-to-end:

  • Customer research & validation

  • Product strategy & experience design

  • UX architecture & flows

  • UI & interaction design

  • AI-assisted prototyping & build

  • Product documentation

  • Engineering collaboration

  • Founder-level decision making

This wasn’t a handoff role — I was embedded in every major product decision.

The Problem We Discovered

Our Product Goal

Create one platform that takes a business from
new lead → paid customer → completed work
without switching tools.

That became our North Star.

Our Opportunity

They didn’t want more software. They wanted one system that worked like their business.

My Process

Using:

  • Customer interviews

  • Founder alignment

  • Rapid prototyping

  • Live production testing

  • AI-accelerated iteration

Discover: Understanding Real Needs

I ran:

  • 9 user interviews by recruiting from local coffee shops, Facebook groups and Sub Reddit forum.

  • Workflow walkthroughs

  • Founder-led sales calls

Using ChatGPT and Gemini, I:

  • Generated interview scripts

  • Synthesized insights

  • Clustered pain points

Key insight:

The biggest problem wasn’t features — it was fragmentation.

Through interviews with real coffee shop owners from actual brick and mortar shops, Facebook groups, and Reddit Forums we learned a sea of common concerns with Inventory, COGs, Employee management and the need of smart AI analytics for forecasting. The POS systems they were paying a lot of money for weren’t giving them the right support, rather just taking % of their profits. They didn’t lack tools. They were drowning in them.

They were juggling:

  • Lead forms

  • Contracts

  • Invoices

  • Payments

  • Job tracking

  • Email

Which led to:

  • Lost leads

  • Missed payments

  • Broken handoffs

  • No single source of truth

Develop: Designing the System

Instead of designing pages, I designed how the business actually worked.

I created:

  • User journeys

  • Data models

  • State transitions

  • Permission rules

  • End-to-end flows

Using:

  • Figma

  • FigJam

  • AI-generated flow modeling

  • Live validation with engineering

Define: Turning Insights into a Product

From research, I defined four product principles:

  1. One source of truth

  2. Minimal setup

  3. Founder-level control

  4. Automation over admin

We mapped the core system:

  • Leads

  • Customers

  • Jobs

  • Payments

  • Documents

This became the foundation of every design and engineering decision.

AI-Powered Prototyping

Cursor AI

To turn designs into working React components and for more robust development work. More importantly bugs and error handling. I was able to fully complete and get it to the 100% production ready level.

Lovable

I used Lovable all the way to 75% for explorative quick design mocks to get to MVP. Which allowed me to collaborate and share the concepts extremely fast with teams and initial user research.

ChatGPT & Gemini

To build the PRD in a very robust way, seek AI for areas that we’ve missed and didn’t think of. Expounded our ideas to more and then narrow towards a tighter direction. Used Gemini to see how it would compare to ChatGPT to avoid mistakes and hallucinations.

Deliver: From Design to Live SaaS in 5 months instead of 12.

I didn’t stop at mockups.

I worked side-by-side with engineering to ship:

  • Onboarding

  • CRM

  • Job tracking

  • Invoicing

  • Payments

And I wrote:

  • Product docs

  • UX specs

  • Feature definitions

  • Flow diagrams

So the team could scale without chaos.

Working with Founders & Engineering

I partnered directly with:

  • Founders on vision, roadmap, and tradeoffs

  • Engineering on feasibility, architecture, and delivery

I ran:

  • Weekly product reviews

  • UX critiques

  • Research readouts

I wasn’t executing orders — I was shaping the product.

Why This Matters

This project shows I can:

  • Build from zero

  • Lead without a PM

  • Design systems, not screens

  • Work with engineers

  • Move fast with AI

  • Stay grounded in real users

This wasn’t a redesign. This was product creation.

Outcome

TassaHub launched as a fully functioning SaaS platform.

We went from:

“We have an idea” → “We have customers using this”

And I delivered:

  • A real product

  • A real system

  • A real business foundation